[clang] a1d57fc - [OpenMP] Do not use the default pipeline without optimizations

[Joseph Huber via cfe-commits]

Author: Joseph Huber
Date: 2022-04-11T17:27:38-04:00
New Revision: a1d57fc22587173c928ed5d82bc6b182d2465eb7

URL: https://github.com/llvm/llvm-project/commit/a1d57fc22587173c928ed5d82bc6b182d2465eb7
DIFF: https://github.com/llvm/llvm-project/commit/a1d57fc22587173c928ed5d82bc6b182d2465eb7.diff

LOG: [OpenMP] Do not use the default pipeline without optimizations

Summary:
A previous patch added the option to use the default pipeline when
perfomring LTO rather than the regular LTO pipeline. This greatly
improved performance regressions we were observing with the LTO
pipeline. However, this should not be used if the user explicitly
disables optimizations as the default pipeline expects some
optimizatoins to be perfomed.

Added: 
    

Modified: 
    clang/tools/clang-linker-wrapper/ClangLinkerWrapper.cpp

Removed: 
    


################################################################################
diff  --git a/clang/tools/clang-linker-wrapper/ClangLinkerWrapper.cpp b/clang/tools/clang-linker-wrapper/ClangLinkerWrapper.cpp
index 618b181e92b53..2dc93ae0988e4 100644
--- a/clang/tools/clang-linker-wrapper/ClangLinkerWrapper.cpp
+++ b/clang/tools/clang-linker-wrapper/ClangLinkerWrapper.cpp
@@ -860,7 +860,6 @@ std::unique_ptr<lto::LTO> createLTO(
   // TODO: Handle index-only thin-LTO
   Backend =
       lto::createInProcessThinBackend(llvm::heavyweight_hardware_concurrency());
-  Conf.UseDefaultPipeline = true;
 
   Conf.CPU = Arch.str();
   Conf.Options = codegen::InitTargetOptionsFromCodeGenFlags(TheTriple);
@@ -868,6 +867,8 @@ std::unique_ptr<lto::LTO> createLTO(
   Conf.MAttrs = getTargetFeatures(TheTriple);
   Conf.CGOptLevel = getCGOptLevel(OptLevel[1] - '0');
   Conf.OptLevel = OptLevel[1] - '0';
+  if (Conf.OptLevel > 0)
+    Conf.UseDefaultPipeline = true;
   Conf.DefaultTriple = TheTriple.getTriple();
   Conf.DiagHandler = diagnosticHandler;